Skip to main content

Hand Detection

MediaPipe provides real-time hand detection and tracking, identifying 21 key landmarks per hand. This section demonstrates hand detection on NVIDIA Jetson using USB and CSI cameras.

1. Hand Detection Overview​

MediaPipe Hand Detection can detect and track hands accurately in real time, making it suitable for: - Gesture recognition - Human--computer interaction - Robotics control - Vision-based interfaces

2. USB Camera Hand Detection​

cd ~/mediapipe

Run USB Camera Script​

python3 01.hand_usb.py
note

Click the preview window and press q to exit the program.

3. CSI Camera Hand Detection​

cd ~/mediapipe

Run CSI Camera Script​

python3 01.hand_csi.py
note

Click the preview window and press q to exit the program.

4. Key Features​

  • Detects 21 hand landmarks
  • Real-time performance on Jetson
  • Supports USB and CSI cameras
  • Lightweight and efficient

Summary​

  • MediaPipe enables accurate hand detection on Jetson
  • Works with both USB and CSI cameras
  • Ideal for interactive and control-based applications

Maintained by HemiHex for Jetson-based advanced vision workflows.